the church in medieval Europe, as well as one example from seventeenthcentury France. The discussion initiates our inquiry into the nature of perpetually lived organizations, an organizational innovation critical for our understanding of open access orders and the transition.

2.2 Commonalities: Characteristics of Limited Access Orders

All natural states share common characteristics. We are ultimately interested in understanding how societies develop the capacity to sustain impersonal rights. Open access orders require that a significant part of the population be treated equally, which necessarily involves treating everyone impersonally without regard to their identity as individuals. To understand how impersonal social relationships evolve we must first understand how larger societies develop that can sustain personal relationships and how personal relationships tie elites to the dominant coalition. These questions lead us into considerations of personality, social networks, and the manipulation of interests. We follow with a few technical notes about the size of a natural state’s coalition and the extent of trade, specialization, and division of labor.

2.2.1 Persons, Personality, Impersonality, Identity,

Patronage, and Interest

In foraging societies, face-to-face interaction among individuals in small groups created personal knowledge, trust (or distrust), and coordination. The limited access order builds on personal relationships and repeated interaction: a hierarchy of personal relationships among powerful individuals at the top of the social order. However, in larger societies, individual relationships cannot be based solely on personal knowledge and trust; they must be reinforced by the web of interests created by the social order. To create stable relationships, individuals must know with whom they are dealing, even if they do not know each other personally.

Societies do not jump directly from personal to impersonal relationships; rather, it is a long process of development that begins in a natural state. On one end of the spectrum, personal relationships are characterized by repeated and idiosyncratic interactions, whereas on the other end of the spectrum impersonal relations are characterized by intermittent and standardized interactions. In other words, all personal relationships are, in some way, unique while large classes of impersonal relationships are the same. In between the two extremes are relationships where the identity of the individuals is uniquely defined, but regularities in interactions between individuals arise.

What then is a person? Every person has two parts. Individual attributes make up one part of every person, including physical size, stature, and appearance, as well as the less tangible characteristics of intelligence, attitude, industry, and ability. The socially ascribed attributes of position, power, privileges, rights, and duties make up the other part of every person. Who we are combines these internal and external parts of our personality. Society recognizes both aspects of personality. When someone is named chief, elected class president, or appointed as department chair, the nature of his or her interpersonal relationships changes: his or her individual identity remains unchanged but his or her social identity changes.

By our nature, each individual’s internal characteristics are unique. However, an individual’s external social characteristics may be unique or those characteristics may be shared with a large group of other individuals. In modern open access societies, for example, the external characteristics of citizens are defined in impersonal terms as a set of social characteristics that apply to everyone who meets certain objective criteria.

Formally, we define a person as composing two interrelated parts: an internal individual persona and an external social persona. The development of impersonal relationships has to do with social persona. As long as social personas are unique across individuals, impersonal relationships are impossible. Impersonality arises as social personas become standardized.

Personality is more complicated than these two aspects. As far back as Roman times, for example, Western law recognized a legal person as any entity capable of bearing rights and duties. What makes an entity a legal person depends on who or what the law decides can bear rights and duties. Slaves and children were not legal persons according to Roman law. A legal person need not be a human being. An incorporeal entity, such as a town or a church, is capable of bearing rights and duties and thus is a legal person. Organizations, therefore, can be legal persons under the right conditions.2 The personality of an organization is always a social persona, defined and supported by the larger society.

In natural states, most relationships within the dominant coalition are personal rather than impersonal. Status and hierarchy tend to be defined in terms of a social persona that is unique to individuals, even if those personas share similarities within broader classes. The notion of nobility in Western

2As Coleman argues in Power and the Structure of Society (1974, pp. 12–13), “In law there are, in fact, two major kinds of persons: physical persons of the sort that you and I know, indeed are what the law calls ‘natural persons,’ and ‘juristic persons.’ The difference between a legal person and a natural person is a staple of legal history.”

Europe is one example. A class of nobles existed. The nobility, however, was further differentiated into dukes, earls, and other grades, and within the grades most nobles possessed unique social identities based in part on the unique privileges each held.

Take a specific but abstract example. The Duke of X possesses certain privileges, rights, and duties that fall on whichever individual corporeal being holds the title, Duke of X. The social persona of the duke includes the formal rights the duke can exercise by virtue of his unique ducal powers and properties. As an individual, the Duke of X possesses his unique features as a human being – his stature, appearance, social grace, ambition, intelligence, focus, and ability to work. His identity as a person encompasses both senses: the duke as a unique individual and the duke as the holder of an office.

The social persona of powerful individuals intertwines inextricably with the organizations that they head or represent. Powerful elites are identified as both individuals and with their organizations. This close relationship between the personal identity of an organization’s leadership and the power of the organization forges the interests that hold natural states together. Powerful members of natural states possess the privilege of forming organizations that the larger society supports and recognizes. How sophisticated and well defined those privileges and obligations are depends on the sophistication of the larger society. Defining the relationship between a person’s individual persona and social persona presents one of the most complicated problems a social order has to solve. The alleged quip of Louis XIV, “L’etat,´ c’est moi,” captures the essence of the problem. Are the official powers of the king vested in the king as an individual or in the king’s social persona, in his dignity, in the office he holds?

In a phrase we will hear again, the question can be framed by asking whether the king is above, below, or identical with the law. Is the ability to change the law the prerogative of the king as an individual person, or is the king bound by the constraints of his social persona? The distinction between the king, duke, pope, or bishop as an individual and the king, duke, pope, or bishop as a social persona became so important in medieval Europe that a formal way of thinking about the individual and corporate aspects of personality developed. Kantorowicz captured this distinction between the individual and social persona in his book, The King’s Two Bodies.3 Whereas the question of a king being above or below the law applies only to the highest levels of society, the question of whether a leader is above or below the organization he heads applies throughout the entire

3The relationship between individual identity and office is treated at some length in Kantorowicz (1997[1957]), Coleman (1974, 1990), and Maitland (2003).

society. If the privilege of using the organization lies with the identity of the individual leader rather than the organization itself, then the society is based on personal relationships. As societies gradually begin to develop ways of privileging organizations as legal persons (entities), irrespective of the personal identity of the leaders, they become capable of sustaining more complicated organizational structures.

As more powerful and sophisticated human organizations, including the state, develop, so does the associated problem of wielding an organization’s power. Is the church’s power at the individual whim of the pope? Does the mayor represent the city? If the mayor errs, must the city pay? These questions are complicated ones, but their answers go straight to our central proposition about social orders: how a social order structures organizations determines the pattern of social interaction within a society. The fact that all organizations must be led by individuals ties the notion of personality and the relationship between individual and social persona directly into the concept of a social order.

The concepts of personality and identity help clarify the position of nonelites in limited access orders and illustrate how the organization of a natural state is reflected throughout the entire social order. The stark way we presented the conceptual framework in the opening chapter may seem to imply that non-elites have no way to obtain or enforce property or security, but some protections for non-elites, their persons and their property, exist in most natural states. Non-elites are not masses of undifferentiated individuals who are treated impersonally. Protection is extended through patronage or clientage networks. The heads of patron–client networks are powerful elites who dispense patronage to clients, provide protection for some aspects of their clients’ property and persons, and negotiate arrangements among elite networks that limit violence if the negotiations are successful. The organization of the network leaders is the dominant coalition of the society. The social identity of non-elites is closely tied to the identity of the patronage network in which they are located: a non-elite is the king’s man or the duke’s man, a Tutsi or Hutu, or any of the millions of group identities that shape human societies.

Natural states include many organizational forms other than patron– client networks, but it may help to think for a moment of natural states as just composed of patron–client types of organizations.4 In natural states, relationships are personal. Nevertheless, because natural states include societies

4Kinship groups, ethnic groups, bureaucracies and other forms of social networks may be organized along patron–client lines, but they need not be. We use patron–client networks as a stand-in for many types of social networks.

with millions of people, personal relationships do not mean that everyone knows everyone else. A hierarchy of elite relationships exists in which small groups of powerful elite individuals know one another through direct personal contact and experience. These circles of elite relationships interlock: all elite individuals know and are associated with other elite individuals above and below them in the social hierarchy. Sometimes elite hierarchies are highly centralized, with a pyramid structure vertically descending from a central king or court. Other natural state hierarchies are much flatter, with more horizontally linked networks of elites.

Natural state elites sit at the top of, but are also embedded in, patron– client networks that extend down into the rest of society. The intra-elite hierarchy roughly corresponds to the hierarchy of the organizations they represent. The most important source of rents binding the interests of the elites together is the rents that flow from their organizations. Patron– client networks combine elements of adherent and contractual organizations. They are simultaneously informal networks grounded on the maintenance of personal relationships between elite leaders and their clients, and more formal organizations where network leaders are able to access thirdparty enforcement of intra-network conflicts by calling on the dominant coalition.

Patron–client networks not only structure the creation, gathering, and distribution of rents that can limit violence; the networks also structure and organize violence itself. When violence breaks out, it is typically among networks of elite factions. Violence works both within and across patron– client networks. The ability of patrons to mobilize their power in aid (or threat) of network members enables them to maintain the network. The patrons’ privileged position within the dominant coalition enables them to protect their clients from injuries caused by clients of other patrons (whether that protection is legal or physical) and their ability to distribute rewards and levy punishments among their clients. The ability of a patron to protect his or her clients depends, in part, on the ability of the patron’s clients to inflict violence on the clients of other patrons. The ability to threaten and use violence is an inherent part of the relationships between elite patrons in the dominant coalition, and between patrons and their clients.

The rewards (rents) for being at the top of the patronage system are typically far higher than those for the patron’s lieutenants, which are again far higher than for the rank and file.5 The rewards of limiting access within the

5Criminal gangs and organizations illustrate the differential returns to members at different levels of the network. Levitt and Dubner (2005) explain why so many drug dealers live

patron–client network create strong incentives for cooperation within the network. Upward mobility within a natural state usually occurs through channels of patronage networks. A talented individual may rise to a position of power through industry and ability. Indeed, a slave in Rome could become a member of the elite. Natural states are limited access societies, not closed access societies. However, the need to rise through patron–client networks ensures that these societies remain limited access social systems.

Natural states include many other forms of organization beyond patron– client networks, and such networks often provide the connections between elites and non-elites. We assume throughout the book that some types of patron–client networks are in place in most natural states.6 From the viewpoint of modern open access societies, patron–client networks appear inherently corrupt. Everything is personal. Whom an individual knows and who they are matter more than what they do. Such an attitude toward natural states, seemingly justified from the viewpoint of an open access society, misses the role that personality, personal relationships, and patronage networks play in containing violence within a natural state: personal relationships and rent-creation provide the incentive systems that contain violence and allow cooperation in a natural state. The inherently personal nature of all relationships in a natural state expresses the fundamental logic underlying the limited access social order.

Personality and identity express themselves in another common aspect of all natural states: the creation and manipulation of interest to ensure social order. If we think of a simple patron–client relationship, where the patron promises to provide protection and the client promises a share or fixed amount of output, the commitments of both the patron and client to

with their mothers: most drug dealers are lieutenants or rank and file, and they cannot afford to live on their own.

6A wide range of excellent case studies of patron–client networks exists. Keefer (2004) and Keefer and Vlaicu (2005) examine patron–client networks in modern developing countries. James Scott’s (1972, 1987) work demonstrates the importance of these networks in Asia for ordering society and providing a modicum of non-elite security and protection. Kettering (1986) describes how elite patronage networks provided an important part of the structure of government in France, a subject we return to later in this chapter. Alston and Ferrie (1985) describe how a patronage network enabled Southern whites to dominate and protect Southern blacks in the late nineteenth and early twentieth centuries, and how those networks were broken down by the expansion of social welfare services in the 1960s. Syme’s (1938) history of the Roman civil wars leading to the end of the Republic focuses on patronage networks as the major unit of analysis. In the next chapter we document the importance of patron–client networks in medieval England. Patron–client networks can be based on kinship, ethnicity, geography, religion, criminal activity, or other factors. They are ubiquitous in natural states.

each other can be credible if they have a long-term relationship. The client can believe the patron will provide protection, because the long-term value of the client’s payments exceeds the cost of protection to the patron, and the patron can believe the client will continue to deliver the payment because the value of protection exceeds the cost of the payment. This requires a personal relationship between the client and patron, a social arrangement in which they are identified with each other. These types of arrangements, throughout society, are more credible and thus easier to sustain if the personal relationships between parties are sustained by social identities that link individuals to each other through organizational ties. Clients have more confidence in patrons if the patrons are embedded in a larger set of social arrangements where the patron’s ability to enjoy a stream of rents from his or her clients is part of what makes the larger arrangement sustainable. In that case, if the patron defaults on his or her clients the patron not only loses the stream of payments from the clients but may also lose the benefits from being part of the larger coalition. The benefits elites receive from heading their networks are part of what make arrangements within the dominant coalition credible, and in turn generate even more benefits for elites. All this depends on identifying who gets the benefits: limited access identifies privileges, creates rents, and provides credibility to personal relationships throughout the society.7 Natural states create and manipulate interests to ensure social order.

The pervasiveness of natural state limits on the ability to form organizations can take the form of a postulate or prediction:

All natural states limit access to organizational forms.

Similarly, natural states control economic opportunities by controlling the organizations and individuals who trade:

All natural states control trade.8

The creation of interests within natural states extends throughout society, well beyond the range of economics and politics. Many natural states treat

7In contrast, in open access societies long-term relationships are still important, but they are embedded in a social structure in which the social identity of individuals does not matter because all individuals are defined in the same terms (i.e., they possess the same privileges or rights).

8Natural states always control who trades, and may also control the places they trade and the prices at which they trade.